Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
74 commits
Select commit Hold shift + click to select a range
f3e33e3
Added some logging
ddobie Mar 9, 2025
d847044
Added dashboard stuff
ddobie Mar 9, 2025
59c07e3
Fixed manager?
ddobie Mar 9, 2025
e8baebb
Remove base_url
ddobie Mar 9, 2025
b89e996
More dask dashboard fiddling
ddobie Mar 9, 2025
9741322
More faff
ddobie Mar 11, 2025
6b611e1
More fiddling
ddobie Mar 11, 2025
3474499
Explicitly collect garbage
ddobie Mar 11, 2025
bd57bdb
Fixed dodgy client load
ddobie Mar 11, 2025
7c3585c
Added objgraph
ddobie Mar 11, 2025
b876fc7
Update config.py to hopefully handle image ingest.
ddobie Mar 12, 2025
57e80ea
Add more objgraph logging
ddobie Mar 12, 2025
0fa9250
del skyreg
ddobie Mar 12, 2025
56336ee
Improved logging
ddobie Mar 11, 2025
97624d0
forced commit
ddobie Mar 12, 2025
052c1a2
More debugging nonsense
ddobie Mar 12, 2025
fa3b261
More change
ddobie Mar 12, 2025
edb8660
Update poetry
ddobie Mar 12, 2025
f862732
Change how dfs are persisted
ddobie Mar 13, 2025
b16334c
Add more checks
ddobie Mar 13, 2025
2e20733
Commit a bunch of junk
ddobie Mar 25, 2025
bc2dde1
Some minor cleanup
ddobie Mar 25, 2025
9afd18c
Fix
ddobie Mar 25, 2025
aa9a3bd
Remove pympler and objgraph
ddobie Mar 26, 2025
aeba51d
Merge branch 'v2.0' into end-to-end-fix
ddobie Mar 27, 2025
edce71f
Merge remote-tracking branch 'refs/remotes/origin/end-to-end-fix' int…
ddobie Mar 27, 2025
b8af6c8
Cleanup
ddobie Mar 27, 2025
4487413
Revert formatting changes
ddobie Mar 27, 2025
2de9a58
Updated changelog
ddobie Mar 27, 2025
bba06c7
Updated changelog
ddobie Mar 27, 2025
89f4023
Do association upload with one worker
ddobie Mar 27, 2025
0c0e3bd
Fixed broken logging statement
ddobie Mar 27, 2025
653b5cc
Further logging cleanup
ddobie Mar 27, 2025
ded5dbd
Add more logging
ddobie Mar 27, 2025
338fb24
Hardcode association upload batch size
ddobie Mar 28, 2025
8862d96
Bump it to 100k
ddobie Mar 28, 2025
9e8e6b3
Revert to 10k
ddobie Mar 28, 2025
6e91e4a
More tweaks
ddobie Mar 28, 2025
135e7a3
Even more logging
ddobie Mar 28, 2025
c7d005b
Clean up logging order
ddobie Mar 29, 2025
92e7426
More detailed timing
ddobie Mar 29, 2025
734da3d
More tweaks
ddobie Mar 29, 2025
3278903
Persist before mapping partitions
ddobie Mar 29, 2025
25aa3b1
Remove repartition
ddobie Mar 29, 2025
3e5a276
Write associations to file before upload
ddobie Mar 29, 2025
ee18d9b
Missed import
ddobie Mar 29, 2025
10079ad
Test finalise fix
ddobie Mar 30, 2025
ff0d234
Change measurements and association DB indices to UUIDs
mauch Mar 30, 2025
a9c4362
Update serializer
mauch Mar 31, 2025
8d4ce2e
Return ones array for new_high_sigma calcs in order to speed up testing
ddobie Mar 31, 2025
7b3a9ab
Clean up logging
ddobie Mar 31, 2025
1aa2f14
Resolve merge conflicts
ddobie Mar 31, 2025
a04e192
More logging
ddobie Mar 31, 2025
bb98195
nerf forced extraction
ddobie Mar 31, 2025
23d4db6
More faff
ddobie Mar 31, 2025
813961a
merge conflicts
ddobie Apr 1, 2025
ad5c3ae
More tweaks
ddobie Apr 2, 2025
d3c1a53
Switch to p2p shuffle
ddobie Apr 2, 2025
5faa416
Do not persist
ddobie Apr 2, 2025
3e129c7
Compute associations df prior to upload
ddobie Apr 2, 2025
db5dda1
More tweaks
ddobie Apr 2, 2025
a5bde90
Merge conflicts
ddobie May 9, 2025
2f30437
Fixed create_measurements_parquet logging
ddobie May 12, 2025
8e2f0d4
Fix issues relating to measurement UUIDs in website rendering
ddobie May 12, 2025
b260bf3
Add memory limitation option
ddobie May 12, 2025
5077edf
Remove hardcoded forced extraction skip
ddobie May 12, 2025
aa26237
More tweaks
ddobie May 12, 2025
14d6e9c
Missing comma
ddobie May 12, 2025
010f1c7
Fixed naming
ddobie May 12, 2025
2c60f1e
stash stuff
ddobie May 14, 2025
f2e8370
Merge remote-tracking branch 'refs/remotes/origin/end-to-end-fix-plus…
ddobie May 14, 2025
68ba615
Resolve merge conflicts
ddobie May 14, 2025
a4bd645
Resolve conflcits
ddobie May 14, 2025
4e16e5c
Fix dask memory management
ddobie May 14, 2025
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 21 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,12 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),

#### Added

<<<<<<< HEAD
- V2: Added additional logging throughout forced_extraction to help monitor memory usage [#820](https://github.com/askap-vast/vast-pipeline/pull/820)
- V2: Added some logging to pinpoint duplicate source ID error - this is likely unnecessary, but keeping in case the error persists [#820](https://github.com/askap-vast/vast-pipeline/pull/820)
=======

>>>>>>> v2.0
- V2: Migrate pipeline to used a Dask.distributed.LocalCluster throughout [#816](https://github.com/askap-vast/vast-pipeline/pull/816)
- V2: Add Dask.distributed support
- V2: Use `django-postgres-copy` for database uploads [#803](https://github.com/askap-vast/vast-pipeline/pull/803)
Expand All @@ -29,12 +34,21 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
- Updated open_fits to correctly handle NaN-padded compressed images [#837](https://github.com/askap-vast/vast-pipeline/pull/837)
- Upgrade measurements file generation to use dask in order to handle larger runs [#797](https://github.com/askap-vast/vast-pipeline/pull/797)
- Replaced optimise_ints and optimise_floats with single function - optimise_numeric [#797](https://github.com/askap-vast/vast-pipeline/pull/797)
<<<<<<< HEAD
- Updated open_fits to correctly handle NaN-padded compressed images [#837](https://github.com/askap-vast/vast-pipeline/pull/837)
- Upgrade measurements file generation to use dask in order to handle larger runs [#789](https://github.com/askap-vast/vast-pipeline/pull/797)
- Replaced optimise_ints and optimise_floats with single function - optimise_numeric [#789](https://github.com/askap-vast/vast-pipeline/pull/797)
=======
>>>>>>> v2.0
- Upgrade forced_phot dependency to v0.2 and force `use_numba=True` [#788](https://github.com/askap-vast/vast-pipeline/pull/788)
- Remove bad forced photometry fits immediately after calculation, rather than after they've all been compiled into a single dataframe [#788](https://github.com/askap-vast/vast-pipeline/pull/788)
- Optimise associations upload by dropping unnecessary columns prior to a large dataframe merge [#787](https://github.com/askap-vast/vast-pipeline/pull/787)

#### Fixed

- V2: Fixed duplicate source ID error by switching from sources_df.repartition() to sources_df.shuffle() in pipeline.finalise.final_operations [#820](https://github.com/askap-vast/vast-pipeline/pull/820)
- V2: Partial fix for image upload memory leak via garbage collect [#820](https://github.com/askap-vast/vast-pipeline/pull/820)
- V2: Fixed dd.concat memory blow-up in pipeline.forced_extraction.forced_extraction by persisting both dataframes prior [#820](https://github.com/askap-vast/vast-pipeline/pull/820)
- V2: Pair metrics working with V2 dask LocalCluster changes [#817](https://github.com/askap-vast/vast-pipeline/pull/817)
- V2: Fixed missing JS9 overlays on source webpage [#809](https://github.com/askap-vast/vast-pipeline/pull/809)
- V2: Fix bug when deleting source tags from database which are referenced by multiple sources [#803](https://github.com/askap-vast/vast-pipeline/pull/803)
Expand All @@ -58,6 +72,10 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),

#### List of PRs

<<<<<<< HEAD
- [#820](https://github.com/askap-vast/vast-pipeline/pull/820): fix: Fixes for duplicate source ID error and memory leak in forced_extraction. Partial fix for memory leak in image upload.
=======
>>>>>>> v2.0
- [#833](https://github.com/askap-vast/vast-pipeline/pull/833): feat: V2: Limit associations upload to using num_io_workers
- [#829](https://github.com/askap-vast/vast-pipeline/pull/829): feat: V2: Allow user specification of dask dashboard paramters and add some further logging to dask setup
- [#817](https://github.com/askap-vast/vast-pipeline/pull/817): fix: V2: Updates to pairs calculation to make it work with Dask `LocalCluster`.
Expand All @@ -69,7 +87,10 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
- [#801](https://github.com/askap-vast/vast-pipeline/pull/801): docs, dep: v2: Updated docs dependencies and fixed outdated information and various warnings
- [#714](https://github.com/askap-vast/vast-pipeline/pull/714): fix: V2: Correctly calculate new_high_sigma parameter
- [#798](https://github.com/askap-vast/vast-pipeline/pull/798): dep: V2: Update dependencies to deal with dependabot reccomendations and to prepare for V2 upgrades.
<<<<<<< HEAD
=======
- [#841](https://github.com/askap-vast/vast-pipeline/pull/841): fix: Fix incorrect changelog link, fix slow image initialisation
>>>>>>> v2.0
- [#837](https://github.com/askap-vast/vast-pipeline/pull/837): fix: Updated open_fits to correctly handle NaN-padded compressed images
- [#828](https://github.com/askap-vast/vast-pipeline/pull/828): fix: Fix potential memory leak in image header extraction
- [#825](https://github.com/askap-vast/vast-pipeline/pull/825): fix: Check for whether processing parameters exist in config before validating them and updated github actions to use cache@v4
Expand Down
Loading